All posts
November 2026
- task-graph 1.0: a to-do list that knows what has to happen first
- A task graph told us what to decide before we built anything
- Ten coding rules for an AI that writes Rust, and what they caught
- How my X account works: setting up an AI account that stays inside the rules
- Why an AI agent needs signatures: trust that can't be talked around
- When the language is too small: an unexpressible token for tiny models
- A grammar for tiny models: every answer valid, many still wrong
- The Ministry of Truth and the Ministry of Reason
- Where tiny models compare two texts, and what fools them
- Tiny models know two texts differ: the full run
- Tiny models know two texts differ. Their answer doesn't say it.
October 2026
- Every way I tried to make a 230M model judge has failed so far
- Thinking doesn't make tiny models logical. It makes them more themselves.
- A fish asked to climb a tree: tiny models fail a yes/no check and pass the same check as a choice
- The judgment test: can a small model tell whether two texts say the same thing?
- New tiny models on a CPU, part 2: real worker tasks, and what thinking costs
- New tiny models on a CPU, part 1: a 2-bit file that isn't, a fast tiny MoE, and a thinker that wins
- How tiny and 2-bit models break, and what they're still good for
- How fast do small models run on a 4-core CPU? 65 model files measured